
A study investigated the genetic basis of an artificial intelligence (AI) algorithm used for predicting the 5-year risk of atrial fibrillation (AF) using 12-lead ECGs. The researchers applied a validated ECG-AI model to ECG data from 39,986 participants without AF and performed a genome-wide association study (GWAS). They discovered three genetic signals at established AF susceptibility loci, including genes related to the sarcomere and sodium channels. Additionally, they identified two novel loci near other genes. In comparison, a GWAS using a clinical variable model showed a different genetic profile. The findings suggest that the ECG-AI model's predictions are influenced by genetic variations associated with specific biological pathways involved in AF.
Like
Save
Share